Hamid Sadeghi
AI Engineer · Machine Learning, Computer Vision & NLP
Summary
AI Engineer with a PhD in Electrical Engineering and over a decade of hands-on experience in applied machine learning, computer vision, and NLP. Experienced in taking AI models from research to production, from computer vision products to conversational systems.
Experience
AI EngineerMar 2023 – Present
Informatics Services Corporation · Tehran, Iran
- Led design and delivery of a Persian-language banking assistant covering 31 services (balances, transfers, cheques, cards, loans, bill payments) using Rasa and LangGraph.
- Integrated LLM-based intent routing to make Persian conversations more natural and robust.
- Built Persian NLU pipelines: intent classification, entity recognition, and context handling.
- Coordinated the work from design to deployment with other developers, including task distribution.
- Developed a CNN-based signature verification module for scanned document matching.
AI EngineerMay 2014 – Mar 2023
Electronics Science Co. · Tehran, Iran
- Designed and delivered multiple versions of face detection and recognition engines (deep CNNs, metric learning) for image and video analysis.
- Built face liveness/anti-spoofing, face quality, and facial attribute models (age, gender, pose, mask).
- Developed real-time object detection and tracking (YOLO, Kalman filtering) and crowd counting models.
- Applied deep learning to disease detection in medical images, including histopathology and MRI.
- Implemented performance-critical modules in C++, deployed models with software teams, and mentored new AI team members.
- Proposed a faster distance formulation for histogram matching in large-scale face recognition, achieving a 2.5x search speed-up with minimal accuracy loss.
- Built a 3D face synthesis module to augment training data under diverse pose and lighting conditions.
- Used OCR, speech, and translation tools for multilingual workflows.
Skills
Areas: machine learning, deep learning, computer vision, NLP and conversational AI, LLM applications, metric learning, face analysis, detection and tracking, medical image analysis
Frameworks & libraries: PyTorch, TensorFlow, Keras, scikit-learn, OpenCV, Transformers, YOLO, Rasa, LangGraph, EasyOCR, SpeechBrain, DeepSpeech, NumPy, pandas
Languages & tools: Python, C++, MATLAB, Git, Linux
Education
PhD, Electrical Engineering (Electronics)2013 – 2019
Amirkabir University of Technology · Thesis: Person-Independent Facial Expression Recognition in the Wild
MSc, Electrical Engineering (Digital Electronics)2011 – 2013
Amirkabir University of Technology · Thesis: Facial Expression Recognition using Image Texture Description
Selected Publications
- HistNet: Histogram-based CNN with Chi-squared deep metric learning for facial expression recognition, Information Sciences, 2022
- Human vision inspired feature extraction for facial expression recognition, Multimedia Tools and Applications, 2019
- Histogram distance metric learning for facial expression recognition, J. Visual Communication and Image Representation, 2019
- Full list (26 publications): hamidsadeghi68.github.io
Other
- 2019 – 2022: AI Technical Committee member and visiting lecturer, Nextera accelerator, University of Tehran. Designed evaluation protocols for Iran's first face recognition challenge (facecup.ir) and taught an online machine learning course.
- 2013 – 2019: Advised MSc projects in computer vision; teaching assistant for Digital Design and instructor of the Electrical Circuits Lab, Amirkabir University of Technology.
- Certificates: ACM Deep Learning Summer School (2018), Scrum Master (2020), Linux LPIC-1 (2021).